Re: + clocksource-driver-initialize-list-value.patch added to -mm tree

[Daniel Walker <dwalker@mvista.com>]

On Wed, 2007-04-04 at 23:15 +0200, Ingo Molnar wrote:
> * Daniel Walker <dwalker@mvista.com> wrote:
> 
> > > > This is a prerequisite for simplifying the clocksource 
> > > > registration process.
> > > 
> > > why? This patch only pushes some unnecessary code into the 
> > > clocksource drivers:
> > > 
> > > +       .list           = LIST_HEAD_INIT(clocksource_avr32.list),
> > > 
> > > NACK unless you can give an explanation of why this is unavoidable. A 
> > > NULL initializer is just as good as an initialized list entry. (in fact 
> > > it's slightly better because it's in the kernel's BSS)
> > 
> > This is only 1 of 9 patches . The 9th patch requires the .list value 
> > to be initialized .. The description change above was suppose to make 
> > that clearer .. By forcing the .list value to be initialized we can 
> > simplify the clocksource registration .
> 
> but why do you call that a simplification? Remove 5 lines of code from 
> the generic code, by adding +1 line to every clocksource driver, 
> totalling to like +20 lines at the moment?

I guess I don't look at it in terms of lines .. Why do you think
reciting a line count diminishes the "simplification" claim? The 20+
lines that I added to each clocksource don't have a size or runtime
effect ..

Daniel